ElasticSearch master and slaves

I don't usually work much with Elastic, I have been reading about Elastic Configuration. It is my understanding that Elastic is designed to run clustered, where several nodes, at least 3, should be master. And there are no real "slaves" it is handled by Elastic.

In my current team we had some issues with the servers not being able to handle the load, and when we were checking the configuration we had 1 master and 2 data-nodes. Btw we have 4 machines for elastic, we have logstash in its own machine.
We discussed this with the Elastic Team and they said that was the correct way to set it up. But everything I have read says the opposite. There should be 3 master nodes. I think the 2 data nodes is ok.

I am in doubt whether I am misunderstanding what I read, or is it that the person in charge of Elastic is used to other kind of Search Engines and has misconceptions.

The 3 nodes should all be master-eligible but only one will act as master at any point in time.

Thanks Christian for your quick answer!
Would it be too bad if we made the 3 nodes, data nodes? It is my understanding that the master related task we do are low, I am unsure about the shards though.

The default setting is that all node have all roles so that is very common in smaller clusters, so no problem at all.

Cool, thanks a lot! Yes i read it is the default, but they changed it all and I wasn't sure.

That depends if you are referring to the one active master, or if there is indeed only one master node configured out of the 3.

_cat/nodes?v should show you what node is the master, and what are master eligible.

It is the one master configured. But thanks a lot!

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.